论文部分内容阅读
基于随机滤波理论的剩余寿命预测模型是基于状态的维修的重要组成部分.首先根据设备磨损过程,建立了磨损、金属粒子浓度和剩余寿命三者的函数关系.进而针对滤波模型基于油液信息进行预测时的局限性,建立了基于油液浓度梯度特征的滤波模型.此模型无需对监测信息中的换油影响进行线性回归处理,从而减少了误差,并以金属浓度梯度特征建模,完善了状态信息与剩余寿命之间的负相关关系.然后设计了极大似然参数估计方法,在参数估计过程中考虑了截尾数据对估计值的影响.最后以某型自行火炮发动机的油液光谱分析数据为例,实现了发动机的剩余寿命预测,结果表明了该模型的可行性和有效性.
The residual life prediction model based on stochastic filter theory is an important part of state-based maintenance.Firstly, according to the wear process of equipment, the function of wear, the concentration of metal particles and the remaining life are established.Furthermore, This paper established a filtering model based on the characteristics of oil concentration gradient.The model did not need linear regression to the effect of oil change in the monitoring information so as to reduce the error and to model the metal concentration gradient feature, The state information and the remaining life.And then, the maximum likelihood parameter estimation method is designed and the influence of the censored data on the estimation is taken into account in the parameter estimation process.Finally, the oil spectrum of a certain self-propelled gun engine As an example, the remaining life of the engine is predicted. The results show that the model is feasible and effective.